Configuring a Model in Ollama

Model name in the Ollama can be configured in two ways

Prerequisites

Ensure that Ollama is installed and running in the background.
Note:
To verify, check the system taskbar notification area.
Download the build file from Passport Advantage.
Place both the build file and the model file in the same directory.
Example: IBD_V10\bin
Open a command prompt in the folder where the files are located.
Run the following command and ensure that the command runs successfully.
```
ollama create GraniteModel1 -f Modelfile
```
Note:
You can replace GraniteModel1 with any model name of your choice.
Run the following command and confirm that the newly created model appears in the list.
```
ollama list
```
Copy the model name and open IBD.
Go to Preferences➔API Keys.
Paste the model name in the Ollama Model name field.
Click Apply, then Apply and Close.
Run the eglcopilot.exe application and when prompted, allow access for domain networks.
Verify that the server starts successfully.
Note:
Do not close the eglcopilot.exe application, it must run in the background.
In IBD, navigate to: Windows➔Show View➔Others.
Select EGL Code Accelerator and click Open.
EGL Code Accelerator opens successfully.
Select the Local Granite model and start interacting with the EGL Code Accelerator.

Note:
Chat history is stored in the location: C:\Users\<username>\EGL Code Accelerator\IBD_chats.

Note:

If you encounter issues accessing EGL Code Accelerator:

Follow the steps below to configure a model hosted on a centralized Ollama server.

Navigate to the API Keys tab in the Preferences window.
In the Ollama URL and Ollama Model Name fields, enter the centralized ollama url and ollama model name as shown in the image.
Click Apply and Close.
The model is now successfully configured in the centralized ollama.